Experimental Physics and
Industrial Control System

From: Mark Rivers <rivers at cars.uchicago.edu>
Sent: Tuesday, March 22, 2022 2:11 PM
To: tech-talk at aps.anl.gov; Smith, William <william.smith at helmholtz-berlin.de>
Subject: Re: NDCodec Compression Factor

 

Hi,

 

I have been revisiting using the NDCodec plugin in areaDetector. I noticed that when I try to repeat the example shown here https://cars9.uchicago.edu/software/epics/NDPluginCodec.html under “Screen Shots” with the following parameters plugin

 

#NDCodecConfigure(const char *portName, int queueSize, int blockingCallbacks, const char *NDArrayPort, int NDArrayAddr, int maxBuffers, size_t maxMemory, int priority, int stackSize)

 

NDCodecConfigure("CODEC1", 20, 0, "SIM1", 0, 0, 0, 0, 0, 5)

 

With the same settings as is shows in the example:

 

Compressor: Blosc

Blosc Compressor: ZSTD

Blosc Comp. Level: 5

Blosc Shuffle: Bit

Blosc num threads: 6

 

I can only achieve compression of something like 1.3 not 151.35 as is shown in the example. What might I be doing wrong?

 

If I use JPEG compression I can achieve higher compression ratios, e.g with JPEG quality set to 10 I get compression factor of 81.5. With anything other than JPEG, for all combinations of settings I get roughly the same compression factor between 1 and 2.

 

The input image is coming from the sim detector. I am using a 1024x 1024 image of UInt8 which is a single peak centred about (500,500).

 

There is no codec error shown. I get no error in the ioc shell.

 

Many thanks,

 

Will